The role of maternal caffeine consumption in the pathogenesis of adverse pregnancy outcomes is controversial. Several studies have found that women who consume caffeine are at increased risk of spontaneous abortion and fetal growth restriction compared to non-users. However, other equally well-done studies have found no harmful effects of caffeine consumption. In addition, several studies have reported that caffeine is harmful only among women who smoke. All previous studies of this question have relied on maternally-reported caffeine use; no studies have employed a biomarker for caffeine. This project first validated the use of serum caffeine and its metabolites as a marker for caffeine intake, and then studied these serum markers as a risk factor for adverse pregnancy outcome. In the validation study, serum paraxanthine was determined to be an acceptable marker for caffeine intake. As part of this project, the concentration of cotinine, a metabolite of nicotine, was assayed in the serum of approximately 450 of the women and the results compared to their reported cigarette smoking. Women were found to be very honest in reporting whether they smoked, but their serum concentration of cotinine correlated only moderately with the amount smoked. This was the case for two separate populations of women 30 years apart. In the main part of this project, the serum concentration of paraxanthine, caffeine's primary metabolite, was compared between 591 women experiencing a spontaneous abortion and 2558 women with live births who had serum drawn at the same time of pregnancy as the women with spontaneous loss. In addition, the association between reduced fetal growth and third-trimester paraxanthine serum concentrations was evaluated among these 2515 women. The mean concentration of paraxanthine was higher in women experiencing a spontaneous abortion than women experiencing a live birth (752 vs 583 ng/ml). Further analysis revealed that this may be explained by a few women with very high concentrations of paraxanthine. The odds ratio for serum paraxanthine concentration greater than the 95th percentile was 1.86 (p<0.01), but intermediate concentrations were not associated with an elevated risk of spontaneous abortion. These results suggest that moderate caffeine consumption during pregnancy does not increase the risk of spontaneous abortion. Among the control women, 2515 provided a third-trimester serum sample. The mean paraxanthine concentration in this sample was higher among women who subsequently gave birth to a small-for-gestational age (SGA) infant (754 ng/ml) than among women who gave birth to appropriately-grown infants (654 ng/ml, p=0.03). There was no statistically significant association between paraxanthine and SGA birth among non-smokers (p=0.48).